Graph Analysis Tool for Network Management Automation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Modern network management faces challenges due to the absence of a cohesive approach for analyzing and manipulating network topologies and communication graphs, leading to a steep learning curve, increased errors, and inefficiencies.

Innovation Solution

The use of large artificial intelligence models, specifically large language models (LLMs), to generate task-specific code for graph manipulation through natural language queries, addressing challenges of explainability, scalability, and privacy by allowing network operators to inspect the generated code and apply it locally.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple diverse tools and domain-specific languages are used for network management tasks, then task functionality is improved, but the learning curve becomes steeper and errors increase

Engineering Contradiction:

Improvetask functionalityVSAvoidlearning curve

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent applies universality by creating a unified natural language interface that can perform multiple network management tasks (graph analysis, topology manipulation, communication pattern detection) through a single system, eliminating the need for multiple specialized tools and domain-specific languages while maintaining full task functionality

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Productivity

If manual operations are performed using traditional tools, then operational control is maintained, but efficiency decreases and errors increase

Engineering Contradiction:

Improveoperational efficiencyVSAvoiderror rate

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent substitutes manual mechanical operations with automated code generation driven by large language models. The system automatically translates natural language queries into executable code for graph manipulation, replacing manual tool operations while maintaining operational control through explainable generated code that can be reviewed before execution

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by allowing network operators to directly query the system in natural language without needing to learn specialized tools or write complex code themselves. The large language model automatically generates and executes the appropriate code based on the operator's intent, making the system serve the user's needs directly

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250036375A1Graph Analysis and Manipulation

Publication Date: 2025.01.30 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

This patent relates to automating network management. One example includes a graph analysis and manipulation tool configured to receive a natural language prompt relating to a network management activity. The graph analysis and manipulation tool is also configured to access a graph resource and to generate code that addresses the network management activity as a graph manipulation task.